Accessing through the Task class¶
loading image¶
import matplotlib.pyplot as plt
%matplotlib inline
from asterism import data_dir
from asterism.core.image_manager.image import Image
from asterism.analysis_tasks.source_detection.background_estimation.background_estimation import DoSetBkgThreshTask
from asterism.analysis_tasks.source_detection.image_segmentation.image_segmentation import DoImageSegmentationTask
from asterism.analysis_tasks.source_detection.deblending.denclue_deblending import DoDENCLUEDeblendingTask
import numpy as np
image=Image.from_fits_file(data_dir+'/deblending_img1.fits')
bkg estimation¶
bkg_task=DoSetBkgThreshTask()
bkg_task.list_parameters()
bkg_task.set_par('plot',value=False)
bkg_task.set_par('skewness_range',value=0.1)
bkg_task.set_par('sub_block_frac_size',value=0.1)
|-------------------------------------------------------
|task: bkg_treshold
|
| par_name= sub_block_frac_size value= None doc= if provided, the bkg pixels are extracted from the sub-block with lowest integrated flux, having a size equal
to sub_block_frac_size time the image size
| par_name= min_size_ratio value= 0.1 doc= provides the minimum size to stop the clipping in the skewness minimization
| par_name= n_std_th value= 4.5 doc= number of bakground deviation to set threshold, th=bkg_level+n_std_th*bkg_sig
| par_name= resampling_size value= 1000000 doc= if the image pixels number>resampling_size,
then a random resampling of pixels is performed, with a final number= resampling_size
| par_name= skewness_range value= 0.1 doc= if not None, |skewness| is minimized in the range |bkg_lvl,bgk_lvl+skewness_range*bkg_std|
| par_name= use_fixed_th value= None doc= if not None, this value is used as th
| par_name= plot value= False doc=
| par_name= verbose value= False doc= set verbosity on
|-------------------------------------------------------
bkg_threshold,bkg_mode,bkg_sig=bkg_task.run(image=image)
|------------------------------
| Task: bkg_treshold Start
|------------------------------
|setting threshold
|image flux min 0.0 image flux max 9.71633911133
|using sub blocks
|sub blocks grid built
|sub block side 31
|sub blocks Num 9
|-
|skewness minimization 961
|flux bkg array min,max 0.0886840820312 0.458374023438
|minimization of |skewness|
|initial skewness 0.160161054539
|initial mean 0.250037394255
|initial std 0.0593948364258
|range of threshold for skewness minimization th_skew in [0.244098,0.255977]
|start size: 961 0.250037394255 0.0593948364258 0.160161054539 961
|min skewness=0.000577 for th_skew=0.062649
|size ratio=0.706556, min_size ratio=0.100000
|-
|bkg_lvl 0.248482923269
|bkg_std 0.0411071777344
|bkg_th 0.433465223074
image segmenetation¶
image_seg_task=DoImageSegmentationTask()
image_seg_task.list_parameters()
|-------------------------------------------------------
|task: image_segmentation
|
| par_name= method value= dbscan_binned doc= method: dbscan or dbscan_binned, from_seg_map
| par_name= bkg_threshold value= None doc=
| par_name= K value= 1.5 doc= sets the segmentation threshold as K*bkg_threshold
| par_name= K_pix value= False doc= if True, total flux in the dbscan bursh is divided by the number of pixels in the brush, and the compared to K*bkg_threshold
| par_name= seg_map_bkg_val value= 0 doc= value for the bkg pixels in the seg map (default 0)
| par_name= dbscan_eps value= 1.0 doc= the scanning brush radius for the dbscan. (In the case of binned, it is the size of the scanning box)
| par_name= verbose value= False doc= set verbosity on
| par_name= plot value= False doc= plot the segmentation results
| par_name= max_image_size value= 1000000 doc= if is not None, the image is segmented in substamps, according to bkg_threshold
| par_name= min_sub_stamp_size value= 25 doc= the minimum substamp size, if smaller the sub stamp will be rejected
| par_name= use_only_central_sub_stamp value= False doc= if True, only the central source is returned
|-------------------------------------------------------
image_seg_task.set_par('bkg_threshold',value=float(bkg_threshold))
image_seg_task.set_par('K',value=1.5)
image_seg_task.set_par('K_pix',value=True)
image_seg_task.set_par('plot',value=True)
image_seg_task.list_parameters()
clusters_list,K,selected_coords=image_seg_task.run(image=image)
|-------------------------------------------------------
|task: image_segmentation
|
| par_name= method value= dbscan_binned doc= method: dbscan or dbscan_binned, from_seg_map
| par_name= bkg_threshold value= 0.433465223074 doc=
| par_name= K value= 1.5 doc= sets the segmentation threshold as K*bkg_threshold
| par_name= K_pix value= True doc= if True, total flux in the dbscan bursh is divided by the number of pixels in the brush, and the compared to K*bkg_threshold
| par_name= seg_map_bkg_val value= 0 doc= value for the bkg pixels in the seg map (default 0)
| par_name= dbscan_eps value= 1.0 doc= the scanning brush radius for the dbscan. (In the case of binned, it is the size of the scanning box)
| par_name= verbose value= False doc= set verbosity on
| par_name= plot value= True doc= plot the segmentation results
| par_name= max_image_size value= 1000000 doc= if is not None, the image is segmented in substamps, according to bkg_threshold
| par_name= min_sub_stamp_size value= 25 doc= the minimum substamp size, if smaller the sub stamp will be rejected
| par_name= use_only_central_sub_stamp value= False doc= if True, only the central source is returned
|-------------------------------------------------------
|------------------------------
| Task: image_segmentation Start
|------------------------------
|method dbscan_binned
|K, 0.65019783461
|bkg th 0.433465223074
denclue deblending¶
deblendig=DoDENCLUEDeblendingTask()
deblendig.list_parameters()
|-------------------------------------------------------
|task: denclue_deblending
|
| par_name= gl_downsampling value= False doc= if True, a GuassLaplace downsampling of the target coordiantes is performed
| par_name= h_grid value= 3 doc= grid mesh in pixels for the downsampling
| par_name= gl_th_rel value= 0.1 doc= relative threshold for the GaussLaplace local maxima detection
| par_name= eps value= 0.01 doc= sets the stop threshold for the recursive attractor update rule
| par_name= digitize_attractors value= False doc= if True, will approximate the final attractor position to the coordinate of the closest image pixel
| par_name= kernel value= gauss doc= smoothing kernel
| par_name= k_table_size value= None doc= if not None, the kernel is precomputed in a look-up table over a grid with size equal to k_table_size
| par_name= h_frac value= 0.1 doc= sets the width of the kernel as h=h_frac*sqrt(r_max^2+r_cluster^2)
| par_name= h_min value= 1.0 doc= kernel width obtained by h_frac can not be lower than h_min
| par_name= h_max value= None doc= kernel width obtained by h_frac can not be lerger than h_max
| par_name= mask_unchanged value= None doc= th value to flag and attractor as unchanged
| par_name= min_size value= 9 doc= sets the minimum size in pixels to perform a deblending
| par_name= attr_dbs_eps value= 1.0 doc= eps for dbscan of attractors
| par_name= attr_dbs_K value= 4.0 doc= K for dbscan of attractors
| par_name= verbose value= False doc= set verbosity on
| par_name= plot value= False doc= plot
+--mex group: R_max_method
x par_name= R_max_frac value= None doc= max kernel influence radius as fraction of parent cluster r_max
x par_name= R_max_kern_th value= None doc= max kernel influence radius when kernel<=th with th relative to kenerl max value
|-------------------------------------------------------
deblendig.set_par('gl_downsampling',value=False)
deblendig.set_par('h_frac',value=0.1)
deblendig.set_par('plot',value=True)
deblendig.set_par('verbose',value=True)
deblendig.run(clusters_list=clusters_list)
|------------------------------
| Task: denclue_deblending Start
|------------------------------
|-
|debelnding for cluster 0 r_cluter 8.24967991157 r_max 15.2908684383 size 378 9
|eps 0.01
|h 1.73743453471
|setting R_max with method auto
|R_max [ 8.35672967]
|running denclue with
|kernel gauss
|h 1.73743453471
|eps 0.01
|R_max set to [ 8.35672967]
| find attractors
|position array size 378
|target position array size 378
|comp. time 0.30864405632
|denclue done
|dbs cl=0, x_c=15.348856, y_c=11.986082
|dbs cl=1, x_c=6.044407, y_c=19.809481
|-
|-
|debelnding for cluster 1 r_cluter 4.07843469095 r_max 7.0304016584 size 77 9
|eps 0.01
|h 1.0
|setting R_max with method auto
|R_max [ 4.80980981]
|running denclue with
|kernel gauss
|h 1.0
|eps 0.01
|R_max set to [ 4.80980981]
| find attractors
|position array size 77
|target position array size 77
|comp. time 0.16042304039
|denclue done
|dbs cl=0, x_c=6.771844, y_c=7.129548
|-
|-
|debelnding for cluster 2 r_cluter 1.34960311626 r_max 1.77658380044 size 8 9
|-
[<asterism.analysis_tasks.source_detection.deblending.deblending.DeblendedProducts at 0x117d77710>,
<asterism.analysis_tasks.source_detection.deblending.deblending.DeblendedProducts at 0x1179eb250>,
<asterism.analysis_tasks.source_detection.deblending.deblending.DeblendedProducts at 0x117e3f710>]
Accessing directly the do_denclue_deblending()¶
from asterism import data_dir
from asterism.core.image_manager.image import Image
from asterism.analysis_tasks.source_detection.background_estimation.background_estimation import set_thresh
from asterism.analysis_tasks.source_detection.image_segmentation.image_segmentation import do_image_segmentation
from asterism.analysis_tasks.source_detection.deblending.denclue_deblending import do_denclue_deblending
image=Image.from_fits_file(data_dir+'/deblending_img1.fits')
bkg_threshold,bkg_mode,bkg_sig=set_thresh(image,sub_block_frac_size=0.1,skewness_range=0.1,plot=False)
|setting threshold
|image flux min 0.0 image flux max 9.71633911133
|using sub blocks
|sub blocks grid built
|sub block side 31
|sub blocks Num 9
|-
|skewness minimization 961
|flux bkg array min,max 0.0886840820312 0.458374023438
|minimization of |skewness|
|initial skewness 0.160161054539
|initial mean 0.250037394255
|initial std 0.0593948364258
|range of threshold for skewness minimization th_skew in [0.244098,0.255977]
|start size: 961 0.250037394255 0.0593948364258 0.160161054539 961
|min skewness=0.000577 for th_skew=0.062649
|size ratio=0.706556, min_size ratio=0.100000
|-
|bkg_lvl 0.248482923269
|bkg_std 0.0411071777344
|bkg_th 0.433465223074
clusters_list,K,selected_coords=do_image_segmentation(image,bkg_threshold,K=np.float_(1.5),K_pix=True,plot=True)
|method dbscan_binned
|K, 0.65019783461
|bkg th 0.433465223074
denclue deblendings¶
do_denclue_deblending(clusters_list=clusters_list,gl_downsampling=True,h_frac=0.1,plot=True)
|-
|debelnding for cluster 0 r_cluter 8.24967991157 r_max 15.2908684383 size 378 9
|R_max 8.24967991157
|comp. time 0.10226392746
|denclue done
|dbs cl=0, x_c=14.810205, y_c=11.953806
|dbs cl=1, x_c=6.911037, y_c=19.636510
|-
|-
|debelnding for cluster 1 r_cluter 4.07843469095 r_max 7.0304016584 size 77 9
|R_max 4.07843469095
|comp. time 0.0175330638885
|denclue done
|dbs cl=0, x_c=8.667780, y_c=2.680334
|dbs cl=1, x_c=7.847731, y_c=5.755992
|dbs cl=2, x_c=8.474922, y_c=7.924534
|dbs cl=3, x_c=4.974973, y_c=8.548267
|dbs cl=4, x_c=6.149232, y_c=9.844471
|-
|-
|debelnding for cluster 2 r_cluter 1.34960311626 r_max 1.77658380044 size 8 9
|-
[<asterism.analysis_tasks.source_detection.deblending.deblending.DeblendedProducts at 0x116124c10>,
<asterism.analysis_tasks.source_detection.deblending.deblending.DeblendedProducts at 0x117730fd0>,
<asterism.analysis_tasks.source_detection.deblending.deblending.DeblendedProducts at 0x11896ffd0>]
